Humoral immune responses to Neisseria are biased toward IgA versus IgG in patients with IgAN.

Based on the observed differences in tonsillar microbial abundance, we hypothesized that patients may exhibit altered titers of anticommensal antibodies in their circulation. Because Neisseria represented the most abundant genus in patients, we measured anti-Neisseria humoral responses, evaluating the ratio of anti-Neisseria IgA to anti-Neisseria IgG in serum. This ratio is reported as an average IgA/IgG response against a panel of 4 commensal Neisseria species (anti–N. lactamica, N. sicca, N. cinerea, and N. flavescens) and 4 pathogenic N. meningitidis (Nme) strains (90/18311, H4476, 208, and 860800). As illustrated in Figure 2, patients with IgAN exhibited an increase in the ratio of plasma IgA/IgG anti-Neisseria antibody titers. Increased IgA titers against Neisseria were observed in response to both pathogenic Nme and commensal strains (Supplemental Figure 3). Therefore, patients with IgAN exhibited a bias toward generating an enhanced IgA response to a variety of commensal Neisseria species and Nme strains compared with healthy controls.

Open in a separate window

Figure 2

The anti-Neisseria response in plasma of patients with IgA nephropathy.

Patients with IgAN exhibited exaggerated IgA-biased anti-Neisseria responses to both pathogenic and nonpathogenic Neisseria species. Neisseria-specific antibodies were evaluated across a panel of 4 N. meningitidis (Nme) strains (90/18311, H4476, 208, and 860800) and 4 commensal strains (N. lactamica, N. flavescens, N. sicca, and N. cinerea). Mean and SD shown, Mann-Whitney U test, 2-tailed P value).

Effect of BAFF overexpression on the immune response to Nme in BAFF-transgenic mice.

We previously reported that the cytokine APRIL (TNFSF13) was elevated in the serum of patients with IgAN compared with controls in 2 independent cohorts (8). We confirmed that APRIL was elevated in patients with IgAN in the current cohort (median 1.98 vs. 1.55; IQR 1.75, 0.38; P < 0.01). Moreover, we observed a positive correlation between serum APRIL levels and proteinuria (Spearman’s ρ = 0.28, P = 0.01). As in our previous study, we did not observe differences in serum BAFF levels.

APRIL binds to TACI, a TNF receptor that promotes class switch to IgA (10). TACI is also stimulated by higher-order multimers of BAFF, and overexpression of BAFF in BAFF-transgenic (BAFF-Tg) mice is sufficient to stimulate TACI, thus mimicking an APRIL/TACI signal (11). Importantly, as they age, BAFF-Tg mice exhibit an IgAN-like disease characterized by underglycosylated IgA in the serum, IgA deposition in the kidney, elevated proteinuria, and kidney pathology (8). We previously showed that the microbiota is an essential cofactor promoting the deposition of IgA immune complexes in the kidneys of BAFF overexpressing BAFF-Tg mice (8). This implies that a host-microbiome interaction may play a role in the etiology of IgAN, at least in mice. However, in our previous studies, we did not assess whether specific microbial candidates could accelerate the pathogenesis of IgAN-like disease in BAFF-Tg mice.

Therefore, we next explored the possibility that Neisseira-targeted, IgA-biased immune responses may be observed in BAFF-overexpressing mice. Nme infects epithelium via the CEACAM-1 receptor. Because Nme will bind to human but not mouse CEACAM-1 we crossed BAFF-Tg mice with mice that express the human form of CEACAM-1 (hC-Tg) to generate double transgenic mice (B × hC-Tg). Transgenic mice and their littermate controls were nasally infected with Nme, and serum and nasal lavages were subsequently collected for analysis. As illustrated in Figure 3A, although hC-Tg^neg (WT) mice were not appreciably colonized by Nme, hC-Tg and B × hC-Tg mice exhibited similar rates of Nme colonization (80% vs. 100%, P = NS) and similar bacterial burdens (192.9 vs. 116.7, P = NS) at 5 days after infection. These data indicate that B × hC-Tg mice did not have altered rates of Nme colonization in comparison to WT mice.

Open in a separate window

Figure 3

The effect of BAFF overexpression on sterilizing immunity against N. meningitidis 90/18311.

(A) Rate of Nme colonization according to genotype during naive infection. (B) Bacterial burden 24 hours after secondary infection in littermate controls. Naive hC-Tg mice were used as a comparison for bacterial burden. (C) At 24 hours after nasal infection, B × hC-Tg mice did not exhibit an exaggerated anti-Neisseria IgA response in the nasopharynx. No anti-Neisseria IgG was detected in mice (not shown). (D) After nasal infection, hC-Tg mice exhibited enhanced systemic anti-Nme IgG response but B × hC-Tg mice did not (1-way ANOVA F 5.0, P < 0.01, adjusted P < 0.05 for indicated comparisons). (E) B × hC-Tg mice exhibited an enhanced systemic anti-Nme IgA response with 10-fold anti-Nme IgA production (1-way ANOVA F 2.6, P = 0.07). (F) Ratio of anti-Nme IgA/IgG revealed IgA-biased systemic response in the B × hC-Tg mice (1-way ANOVA F 2.3, P = 0.09, unadjusted P value as shown). All tests 1-way ANOVA with Tukey’s test; adjusted P value, mean, and SD shown.

We have previously shown that 2 nasal infections are required to induce immunity against a subsequent third nasal infection in hC-Tg mice (12). To determine whether B × hC-Tg mice exhibited improved clearance of Nme compared with hC-Tg controls, we nasally infected mice twice, on days 1 and 14. At 24 hours after secondary infection, the rate of infection of hC-Tg mice and B × hC-Tg mice was similar (85% vs. 80%), demonstrating a lack of neutralizing immunity induced by 1 previous exposure of Nme (Figure 3B). This implies that overexpression of BAFF did not eliminate the need for 2 infections to induce protection against colonization in our mouse model.

Given that there was no obvious impact of BAFF overexpression on nasal susceptibility to Nme, we looked at the mouse antibody response to Nme colonization. First, we assessed the local IgA and IgG response to Nme. Nasal Nme-specific IgG, as evaluated in nasopharyngeal lavage fluid, was not detected in any mice, and nasal anti-Nme IgA titers were similar for both B × hC-Tg and hC-Tg mice (Figure 3C). We next measured the systemic antibody response to Nme infection. After the second nasal infection, both hC-Tg and B × hC-Tg mice exhibited an Nme-specific IgG response in serum above baseline, with hC-Tg mice exhibiting a mean anti-Nme IgG titer 2-fold greater than that in B × hC-Tg mice (164 ng/mL vs. 54.6 ng/mL) (Figure 3D). In contrast, B × hC-Tg mice exhibited a mean anti-Nme IgA titer in serum that was 10-fold higher than in hC-Tg controls (321.7 ng/mL vs. 33.09 ng/mL) (Figure 3E). Taken together, our findings demonstrated that whereas the local IgA response to Nme was comparable between hC-Tg and B × hC-Tg mice, the systemic humoral immune response to nasal Nme exposure was IgA-biased in B × hC-Tg mice (Figure 3F).

Microbiota characterization.

Tonsil bacteria were sampled using 2 sterile specimen collection swabs, which were placed immediately into cryovials and stored at –80°C. The samples were suspended in PBS and proteinase K and disrupted by vortex. Bacterial DNA was extracted using the DNeasy Blood and Tissue column-based kit (Qiagen). The DNA integrity and quantity were evaluated using a Nanodrop spectrophotometer (Thermo Fisher Scientific). Individuals were instructed to collect their stool using a standard sterile specimen collection container and then to place it in the home freezer. All stool samples were snap-frozen upon receipt at the research lab. After sample homogenization with 0.1 mm glass beads (MoBio), stool bacterial DNA was extracted using the QIAamp DNA Stool Mini kit (Qiagen) according to the manufacturer’s protocol.

The V4 hypervariable region of the 16S rRNA gene was amplified using a universal forward sequencing primer and a uniquely barcoded reverse sequencing primer to allow for multiplexing (21). Amplification reactions were performed using KAPA2G Robust HotStart ReadyMix (KAPA Biosystems) following the manufacturer’s protocol. For saliva samples, the V4 region was amplified by an initial denaturation at 95°C for 3 minutes, followed by 24 cycles of 95°C for 15 seconds, 50°C for 15 seconds, and 72°C for 15 seconds, ending with a 5-minute extension at 72°C. For stool samples, the amplification reactions were carried out for 17 cycles instead of 22 cycles. All amplification reactions were performed in triplicate, checked on a 1% agarose tris-borate-EDTA electrophoretic gel (TBE) gel, and then pooled at even concentrations. The final library was purified using Agencourt AMPure XP beads (Beckman Coulter). The purified library was quantified and sequenced on the Illumina MiSeq System, according to the manufacturer’s instructions, using the 300 cycle V2 sequencing chemistry to generate 150 × 2 paired-end reads.

Sequencing data were demultiplexed using QIIME2 standard paired-end demultiplexing protocol or directly on the Illumina MiSeq. The UNOISE pipeline, available through USEARCH version 10.0.240, was used for sequence analysis (22–24). The last base, typically error-prone, was removed from all the sequences. Sequences were assembled and quality-trimmed using –fastq_mergepairs and –fastq_filter, with a –fastq_maxee set at 1.0 and 0.5, respectively. Assembled sequences less than 233 bp were removed. Following the UNOISE pipeline, unique sequences were identified from the merged pairs and sorted. Sequences were denoised and chimeras were removed using the unoise3 command in USEARCH. Assembled sequences were then mapped back to the chimera-free denoised sequences at 97% identity operational taxonomic units (OTUs) using the –usearch_global command. Taxonomy assignment was executed using SINTAX (25), available through USEARCH, and the SINTAX compatible Ribosomal Database Project (RDP) database version 16, with the default minimum confidence cutoff of 0.8 (26). OTU sequences were aligned using PyNast accessed through QIIME (27). Sequences that did not align were removed from the data set, and a phylogenetic tree of the filtered aligned sequence data was made using FastTree (28). Low-abundance OTUs (<0.005% RA) were removed from the OTU table (29).

Beta diversity was calculated using QIIME (27). The data were rarefied to an even depth of 10,000 sequences per sample. Principal coordinate analysis plots were of the rarefied data using QIIME and plotted using EMPeror (30).

Preparation of Neisseria cultures.

Neisseria strains were cultured on gonococcus (GC) agar supplemented with IsoVitalex and in the context of mouse experiments, VCNT (vancomycin, colistin, nystatin, and trimethoprim) inhibitor (Becton Dickinson) at 37°C with 5% CO₂. Log phase growth was achieved by transferring overnight cultures from GC agar into 10 mL of brain heart infusion (BHI) broth (Becton Dickinson) supplemented with 1% IsoVitalex and incubated at 37°C with agitation for 4 hours. Broth cultures were diluted to an OD₆₀₀ = 0.2 prior to heat-killing at 65°C for 30 minutes.

Serum procedures.

To measure APRIL in serum samples, we used a commercially available kit (eBioscience, BMS2008). The kit protocol was used with the following modifications: human serum samples were diluted 1:3 in nuclease-free water and subsequently heat-inactivated in a 60°C water bath for 5 minutes. Heat-inactivated samples were then loaded neat onto the microwell strips without using sample diluent. Standard was diluted in the plate, in duplicate, as indicated in the protocol but extended for a total of 13 dilutions followed by a sample diluent blank, a water blank, and an internal serum control sample from a female study participant without IgAN, heat-inactivated in the same manner as the other samples. Samples were incubated without the anti-APRIL detection antibody for 2 hours, as per the protocol. After washing the plates, 50 μL/well of the anti–APRIL-biotin antibody was added for 1 hour and shaken at room temperature, as per the protocol. Lastly, the plates were developed for 12–13 minutes before stopping the reaction and reading on a photospectrometer at 450 nm.

Serum Neisseria-specific IgA and IgG were measured by a custom whole-bacteria ELISA as previously described (12). Briefly, multi-well plates (Nunc Maxisorp, Nalgene) were coated with heat-killed Neisseria strains and left to dry overnight, and then washed and blocked with 5% BSA. Patient plasma samples were diluted 1:1000 and assayed for Neisseria-specific antibodies across a panel of 4 Nme strains (90/18311, H4476, 208, and 860800) and 4 commensal Neisseria strains (N. lactamica, N. flavescens, N. sicca, and N. cinerea). Patient samples and standards were applied to the wells in duplicate and incubated overnight. A standard curve was created using human plasma IgA (Calbiochem). After incubation, plates were washed again and the detection antibody, an alkaline phosphatase–conjugated (AP-conjugated) goat antibody targeting human IgA (Jackson ImmunoResearch), was applied. Absorbance at 600 nm wavelength was measured after 40-minute incubation with BluePhos AP detection substrate (SeraCare, KPL).

Mice.

BAFF-Tg mice (obtained from Ann Ranger and Jeff Browning, Biogen Inc.) were backcrossed with WT mice (Charles River Laboratories) and subsequently interbred as BAFF-Tg^+/+ or BAFF-Tg^+/– mice (7, 8). Mice that express transgenic human CEACAM-1 (hC-Tg) have been previously described (12). These mice were crossed with BAFF-Tg mice to generate BAFF × hCEACAM-Tg progeny (B × hC-Tg) for infection experiments.

Tissue mRNA expression.

At the time of euthanization, a section of kidney tissue was preserved in formalin and paraffin-embedded to be used for morphological assessments and RNA extraction. Tissue was sectioned and collected in 100% xylene for deparaffinization and washed with 75% ethanol. Tissue disruption and protease digestion were performed as previously described (31). The Recoverall kit (Invitrogen) was used for remaining steps of column-based RNA extraction. Reverse transcription was performed using cDNA generated using SuperScript IV Reverse Transcriptase (Invitrogen). Abundance of the secreted-spliced variant of IgA mRNA was quantitated using real-time PCR with custom primers and normalized to GAPDH expression. IgA: forward 5′-GCC TTG CCC ATG AAC TTC AC-3′, reverse 5′-CGC TGA CAT TGG TGG GTT TA-3′; GAPDH: forward 5′-CAT GGC CTT CCG TGT TCC TA-3′, reverse 5′-GCG GCA CGT AG ATC CA-3′. Samples were diluted with RNase free water and reactions contained 5 μL 2× Maxima SYBR Green (Thermo Fisher Scientific), 0.4 μL each of forward and reverse primers, 2.2 μL nuclease-free water, and 2 μL cDNA. All reactions were run in duplicate on a 384-well plate using the Bio-Rad CFX384 Touch Real-Time PCR Detection System.

In vivo Neisseria infections.

WT (hC-Tg^neg), C-Tg, and littermate B × hC-Tg mice were nasally infected with 1 × 10⁵ CFUs of Nme strain 90/18311 resuspended in 10 μL of PBS. Mice were infected at day 1 (primary infection) and, where indicated, day 14 (secondary infection). Mice were euthanized via CO₂ inhalation 5 days after primary infection or 1–4 days following the secondary infection. Serum used to measure systemic responses to Nme was collected via cardiac puncture; mucosal samples were obtained via tracheal/nasal lavages. Bacterial burden per mouse was enumerated by swabbing nasal cavities and plating resuspended fluid overnight at 37°C.

Commensal-specific ELISPOT.

Membrane plates (0.45 μm Hydrophobic High Protein Binding Immobilon-P Membrane, MilliporeSigma) were coated with 5 mg/mL of heat-killed inactivated Nme and placed at 4°C overnight. The plates were blocked the next day with 10% FBS/complete RPMI for at least 2 hours at 37°C. Single-cell suspensions were loaded onto the plate at serial 2-fold dilutions in 10% FBS/complete RPMI and incubated overnight at 37°C. Cells were removed the next day and washed with 0.1% Tween 20/PBS 5×. HRP-conjugated IgA and AP-conjugated IgG (in the case of 2-color ELISPOT) detection antibodies were subsequently added for 2 hours at 37°C.

Plates were then washed with 0.1% Tween 20/PBS 3× and with PBS 3×. The plates were developed while covered with aluminum foil until spots were visible using 3-amino-9-ethylcarbazole (AEC) peroxidase (for HRP-conjugated antibodies, Vector Laboratories) and Vector blue (for AP-conjugated antibodies, Vector Laboratories) substrates. After the development was done, the plates were washed with distilled water and were left to dry overnight. The spots were counted based on the original cell dilution.

Preparation of single-cell suspensions from kidney.

Mouse kidneys were collected in 10% FBS/complete RPMI and kept on ice until digestion was performed. The kidneys were perfused with PBS for the ELISPOT tests. Multi Tissue Dissociation Kit 1 enzymes (Miltenyi Biotec) were prewarmed to 37°C and in accordance with the manufacturer’s recommendations. Both mouse kidneys were incubated with enzymes in a gentleMACS C tube using MACSmix Tube Rotator for 10 minutes at 37°C in a CO₂ incubator. Next, the mixture was briefly transferred to a 6-well plate and mechanically dissociated with tweezers into approximately 1 mm × 1 mm pieces, and then returned to the C tube for an additional 10 minutes of incubation as above. After incubation, a gentle MACS dissociator was used for a final mechanical digestion step prior to filtering through 70 μm nylon mesh and inactivation of enzymes using ice-cold RPMI with 10% FBS.

A Percoll gradient (30% Percoll for 20 minutes at 200 rpm [730 relative centrifugal force × g]) was performed on the cells obtained from the digested kidneys to enrich for immune cells. The pellet was then resuspended in 1 mL of RBC lysis buffer (155 mM NH₄Cl, 12 mM NaHCO₃, 0.1 mM EDTA) for 5 minutes on ice. The cells were washed with PBS 1× and with 10% FBS/complete RPMI 1× and were then ready to be plated.

Statistics.

Data distribution was evaluated and group comparisons were performed using unpaired parametric or nonparametric tests as appropriate. A P value less than 0.05 was considered statistically significant.

HNR’s work is supported by the Gabor Zellerman Chair in nephrology research from the Department of Medicine at the University of Toronto. This project was supported by the Toronto Department of Medicine Challenge grant, the Kidney Foundation of Canada, and the generous support of the Pearson family. The work is also supported by a Canadian Institutes for Health Research operating grant. JLG’s work is supported by a Foundation grant from the Canadian Institutes for Health Research.